How AI is helping New Zealand build safer after disaster

Key Concepts

• New Zealand Geotechnical Database (NZGD): A centralized repository for geotechnical data established by the NZ government.
• BEYON: A digital twin platform developed by Beca to manage, query, and visualize geotechnical data.
• Agentic AI: An AI system configured via Microsoft Foundry that acts as an interface to automate data retrieval and analysis.
• Geotechnical Information: Data regarding soil and rock properties essential for structural engineering and infrastructure design.
• Digital Twin: A virtual representation of physical assets or data environments used for simulation and management.

The Genesis of the New Zealand Geotechnical Database (NZGD)

The 2011 Christchurch earthquake served as a national catalyst, exposing critical vulnerabilities in New Zealand’s infrastructure planning. The disaster highlighted a systemic failure: the lack of centralized, accessible geotechnical data. Engineers struggled to design resilient structures because essential information regarding soil and rock composition was fragmented and uncollated. In response, the New Zealand government established the NZGD to standardize and house this vital information. However, as the database expanded, it faced significant challenges regarding scalability and standardized accessibility.

BEYON: The Digital Twin Solution

Beca identified the need for a more sophisticated management layer to handle the growing complexity of the NZGD. They developed BEYON, a digital twin platform designed to serve as the primary interface for engineers.

• Functionality: BEYON allows users to manage, query, and download complex geotechnical datasets efficiently.
• Purpose: It acts as the bridge between raw, historical data and modern infrastructure development, ensuring that lessons learned from past natural events are integrated into future construction projects.

AI Integration and Operational Efficiency

To further optimize the utility of BEYON, Beca integrated an AI system built on Microsoft Foundry. This "agentic AI" acts as an intelligent intermediary between the engineer and the database.

• Methodology: Instead of manually searching through vast, dense logs of soil and rock data, engineers use the AI to query the database. The AI interprets the request and retrieves only the relevant data points.
• Performance Metrics: The implementation of this AI has resulted in a 40% reduction in time required to access and process geotechnical data.
• User Impact: By automating the filtering of "intense" soil and rock logging records, the AI eliminates the need for engineers to sift through hundreds of irrelevant data points, significantly reducing cognitive load and human error.

Strategic Importance and Future Outlook

The integration of AI into geotechnical workflows represents a shift from reactive data management to proactive, informed engineering.

• Key Argument: The ability to build resilient infrastructure is directly proportional to the quality and accessibility of historical data.
• Supporting Evidence: The 40% efficiency gain demonstrates that technological intervention (AI/Digital Twins) is not just a convenience but a necessity for managing the scale of data required for modern seismic and climate-resilient design.
• Significant Statement: As noted by the user experience, the AI has become an indispensable tool: "I’d be lost without it. I don’t want to know a world without it."

Conclusion

The evolution from the fragmented data landscape of 2011 to the AI-driven BEYON platform marks a significant advancement in New Zealand’s engineering capabilities. By leveraging digital twins and agentic AI, Beca has transformed the NZGD from a static repository into an active, queryable resource. This technological framework ensures that engineers can draw upon historical geotechnical insights with unprecedented speed and precision, ultimately leading to safer and more resilient infrastructure development.